Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The human melanoma-associated antigen p97 is attached to the plasma membrane by a glycosyl-phosphatidylinositol… Food, Michael Robert 1992

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


831-ubc_1992_fall_food_michael.pdf [ 963.35kB ]
JSON: 831-1.0086521.json
JSON-LD: 831-1.0086521-ld.json
RDF/XML (Pretty): 831-1.0086521-rdf.xml
RDF/JSON: 831-1.0086521-rdf.json
Turtle: 831-1.0086521-turtle.txt
N-Triples: 831-1.0086521-rdf-ntriples.txt
Original Record: 831-1.0086521-source.json
Full Text

Full Text

THE HUMAN MELANOMA-ASSOCIATED ANTIGEN p97 Is ATTACHEDTO THE PLASMA MEMBRANE BY A GLYCOSYLPHOSPHATIDYLINOSITOL ANCHORTyMICHAEL ROBERT FOODB.Sc., The University of Victoria, 1986A THESIS SUBMITTED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFMASTER OF SCIENCEinTHE FACULTY OF GRADUATE STUDIES(Biotechnology Laboratory and the Department of Microbiology)We accept this thesis as conformingto the required standardTHE UNIVERSITY OF BRITISH COLUMBIAOctober 1992© Michael Robert Food, 1992In presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Columbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives. It is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.Department of t1ic Oj3IOLDrYThe University of British ColumbiaVancouver, CanadaDate D& i iqqA?,DE-6 (2/88)11ABSTRACT:Melanotransferrin, or p97 is a cell surface glycoprotein which was firstdescribed as a marker antigen for human melanoma cells. Although p97 hasa high level of amino acid sequence homology to human serum transferrinand lactoferrin its function has not yet been determined. One feature thatdistinguishes p97 from the other members of the transferrin family is thepresence of a stretch of 24 hydrophobic amino acids at the C-terminus. In allprevious studies, this sequence was thought to form a proteinacioustransmembrane domain. In this study, however, sensitivity to bacterialphosphatidylinositol-specific phospholipase C, biosynthetic labelling with[3H]-ethanolamine, and partitioning in Triton X-114, were used to establishthat p97 is expressed at the cell surface as a glycosyl-phosphatidylinositolanchored protein. As well, this anchor was shown to be present in p97 onhuman tumor lines and it apparently confers no special intracellulartransport properties on the molecule. These findings raise importantquestions about the function of p97 and the possible involvement of thisprotein in a cellular iron uptake system that is independent from thetransferrin/ transferrin receptor system.111TABLE OF CONTENTS:ABSTCT.iiLIST OF TABLES ivLIST OF FIGURES vACKNOWLEDGEMENT viiLIST OF ABBREVIATIONS viiiINTRODUCTION 1MATERIALS AND METHODS 12RESULTS 20DISCUSSION 50CONCLUDING STATEMENT 57REFERENCES 58ivLIST OF TABLES:Table I: Summary of the properties of the MAbs used in thisstudy 21Table II: Effect of various PT-PLC preparations and pronase on human andmouse cell surface antigens 29Table III: Effect of bacterial PT-PLC and pronase on human and mouse cellsurface antigens 32Table IV: Effect of bacterial PT-PLC on human p97 expressed on the cell surfaceof various CHO lines 39Table V: Effect of BFA on the transport of p97 to the surface of SK-MEL-28cells 46Table VI: Effect of bacterial P1-PLC on human p97 expressed on the cell surfaceof various human intestinal lines 49VLIST OF FIGURES:Figure 1: General structural features of GPI-anchors 8Figure 2: Summary of cDNA constructs used to transfect CHO lines 16Figure 3: Reactivity of various MAbs with surface antigens of SK-MEL-28cells 22Figure 4: Reactivity of various MAbs with surface antigens of p97aTRVBc3cells 23Figure 5: Reactivity of various MAbs with surface antigens of EL-4cells 24Figure 6: Titration of purified L235 MAb reactivity against SK-MEL-28 cells...25Figure 7: Titration of purified C MAb reactivity against SK-MEL-28 cells 26Figure 8: Titration of purified OKT9 MAb reactivity against HuTu-80 cells... .27Figure 9: Effect of bacterial P1-PLC on cell surface antigens 30Figure 10: Phase separation of p97 and TR in Triton X-114 solution 34Figure 11: Summary of the development of the CHO line WTB expressinghuman p97 36viFigure 12: Summary of the development of the CHO line TRVB expressinghuman p97 37Figure 13: Effect of bacterial PT-PLC on human p97 expressed on the cellsurface of various CHO lines 38Figure 14: Effect of bacterial P1-PLC on human p97 expressed on the cellsurface of SK-MEL-28 and p97aWTBc7 lines 40Figure 15: Biosynthetic labelling of p97 with[3H1-ethanolamine in SK-MEL-28cells 42Figure 16: Biosynthetic labelling of p97 with[3H1-ethanolamine inp97aWTBc3 cells 44Figure 17: Biosynthetic labelling of p97 with[3H1-ethanolamine inp97aTRVBc3 cells 45Figure 18: Reactivity of various MAbs with surface antigens of CaCo-2cells 47Figure 19: Reactivity of various MAbs with surface antigens of HuTu-80cells 48viiACKNOWLEDGEMENT:The efforts of Dr. Sylvia Rothenberger and Mr. Ian Haidl in developingand performing the cell surface biotinylation experiments are gratefullyacknowledged. I would also like to thank Dr. W. A. Jefferies for providing theopportunity, facilities, and guidance which enabled me to complete this work.As well, I would like to thank my thesis advisory committee and mycolleagues in the laboratory for their contributions. Finally, I would like tothank the players, training staff, and coaching staff of both UBC andMeralomas Rugby Clubs for allowing me to continue playing rugby duringthese difficult times.viiiLIST OF ABBREVIATIONS:The abbreviations used are: UVR, ultraviolet radiation; Tf, transferrin; TR,transferrin receptor; MAb, monoclonal antibody; MAbs, monoclonalantibodies; GPI, glycosyl-phosphatidylinositol; VSG, variable surfaceglycoprotein; PT-PLC, phosphatidylinositol-specific phospholipase C; BFA,Brefeldin A; ATCC, American Type Culture Collection; DMEM, Dulbecco?sModified Eagle Medium; FBS, Fetal Bovine Serum; CHO, Chinese HamsterOvary; FACS, fluorescence activated cell sorting; SDS-PAGE, sodium dodecylsulfate-polyacrylamide gel electrophoresis; biotin, biotinamidocaproate Nhydroxysuccinimide ester; PMSF, phenylmethylsulfonylfluoride, MEM,Minimum Essential Medium.1INTRODUCTION:The increase in the occurence of skin diseases in many industrializednations has been attributed in large part to the increased exposure of thepopulation to ultraviolet radiation (UVR) present in sunlight (1). Thedamage which results from this exposure to UVR can include: erythema,sunburn, photodamage, photocarcinogenesis, eye damage, immune systemalterations, and chemical hypersensitivity (2). The most common skincancers produced by exposure to UVR are the non-melanoma skin cancers,which are termed basal cell and squamous cell carcinomas (2). These cancershave been associated with repeated sun exposure (3). The less common skincancer, cutaneous malignant melanoma, has been associated with shorter,more intense exposure to the sun (4). This form of skin cancer is both rapidlyincreasing and highly fatal, in fact, it has been suggested that by the year 2000,approaching 1 in a 100 caucasians will develop malignant melanoma, and 20-30% of these people will ultimately die from it (5). The variety of risk factorswhich contribute to the development of malignant melanoma have not beencompletely described.Iron is required by all cells (6), consequently many different systemshave been developed to obtain iron from the environment. Two problemswhich any system must address are that free iron is toxic and that ironchemistry favors the production of insoluble and inaccessible hydroxides (7).Iron is a component of a large number of proteins with essential cellularfunctions, thus, it is absolutely required for cell proliferation (6). Irontransport into microorganisms can be mediated by a high affinity systemconsisting of a soluble iron ligand (siderophore), a membrane bound receptor(ferri-siderophore), and an enzymatic release mechanism, as well as by aniron reductase system (8).2In vertebrates, the iron requirement is thought to be provided by thebinding of iron to the major serum iron transport protein called transferrin(Tf). The Tf molecule is a monomeric glycoprotein of 80,000 daltonsmolecular mass (9). It can reversibly bind two Fe3+ ions per molecule,simultaneously with two C032- ions (9). Evidence from the Tf sequence andproteolytic digestion suggest that the molecule can be subdivided into twohomologous domains of approximately 340 amino acids which share 35-40%sequence identity to one another (9). Each domain has a single iron bindingsite which suggests that the transferrins evolved by gene duplication of asingle iron binding domain (9). The currently accepted view is that soluble Tfbinds iron and this complex interacts with the transferrin receptor (TR) onthe plasma membrane (10). The TR is composed of two identical disulfidelinked polypeptide chains, each with a molecular mass of 95,000 daltons (10).The efficiency of this mechanism is increased because the TR can bind twomolecules of iron-Tf per receptor (6). After binding, the iron-Tf/TR complexremains membrane associated and is concentrated in coated pits andinternalized. The resulting endosomes become acidified and the iron, likelyFe3+, is released from the complex (10). The apotransferrin remains bound tothe receptor and is recycled to the cell surface where it is released at neutralpH and can participate in the binding and uptake of additional iron into thecell (6). The iron in the acidified endosome is transported across theendosomal membrane by an uncharacterized mechanism, where it serves as asubstrate for the biosynthesis of iron containing proteins or is stored inferritin deposits (10). Although cellular iron uptake has been shown to bemediated mainly by the Tf/TR pathway (11), there is evidence for non-Tfmediated pathways of iron incorporation in leukemic cells (12), HeLa cells3(13,14), and in hepatocytes (15). In human melanoma cells, a non-Tfmediated pathway has also been investigated (16-19).Melanotransferrin, also known as the human melanoma-associatedantigen p97’, was one of the first cell surface markers associated with humanskin cancer (20). The p97 molecule was characterized using monoclonalantibodies (MAb) raised against melanoma cells (20-22), including the SKMEL-28 cell line, and p97 has been termed a melanoma specific marker. Itwas shown to be a monomeric membrane bound protein with a molecularmass of 97,000 daltons (23). The p97 molecule was proposed to be both apredictive marker for melanomas, and the target site for immunotherapy(24). Subsequent work with p97 on SK-MEL-28 cells revealed that it wascapable of binding iron (25). Furthermore, p97’, like Tf and the TR, is asialoglycoprotein, which is encoded on chromosome 3 in humans (26). Theprimary structure of p97 deduced from its mRNA sequence (27) indicated thatit belongs to a group of closely related iron binding proteins found invertebrates. This family includes serum Tf, lactoferrin, and avian egg whiteovotransferrin. The human p97 and lactoferrin molecules are known toshare 40% sequence identity (9). In contrast to the other molecules of the Tffamily, p97 is the only one which is directly associated with the cellmembrane, which has implications for the evolution of the Tf family and,potentially, for iron transport mechanisms in general.The p97 molecule was described independently by other investigators.It was shown that p97 was the same as the gp95 molecule (23), described byDippold (28). Subsequently, the gp87 molecule was also shown to be identicalto the p97 molecule (29). More recently, a unique tumor antigen of humanmelanoma was described in detail, and designated gp9O (30). This antigen wasshown to be expressed in high quantities on the surface of a wide variety of4cultured cells, but only in small quantities on cells of a limited range ofnormal tissues (30). It was not until 1989 that it was demonstrated that thisunique antigen was in fact p97 and not a new melanoma-associated antigen(31). One benefit of these studies is that a large number of unique MAb to p97were generated, which collectively defined many different epitopes of themolecule.The distribution of the p97 molecule has been described in considerabledetail over the last twelve years. The first quantitative analysis of p97expression in normal and neoplastic tissues showed that it was present inlarge amounts in most melanomas, some tumors and certain fetal tissues, butin normal tissues it was expressed in small amounts (21). Further work bythe same group again demonstrated that p97 was expressed by cells frommelanoma tumor biopsies but not in normal adult tissues (22). Oneinteresting observation was that p97 was present in high quantities in thefetal colon, lung and umbilical cord (22). In subsequent years the distributionof the p97 molecule continued to be investigated, most notably in the form ofa comparative study of antigen expression in a variety of cell types (32). Thep97 molecule was again shown to be present on the cell surface of a numberof melanoma and other cancer lines, and absent on the cell surfaces of a widevariety of normal adult tissues (32). An exhaustive study of the expression ofthe p97 molecule in cultured tumor and normal cells (30) showed that p97was expressed at a high level in cell lines ranging from melanoma to colon,kidney and lung cancer; moreover, in normal cells, only melanocytes, kidneyepithelium and fibroblasts showed a degree of p97 expression. It has beensuggested that the distribution of expression of the p97 molecule indicatesthat p97 is actually an oncogene.5The situation is not as clear as the majority of these results would seemto indicate. It has subsequently been shown that the p97 molecule isexpressed on endothelial cells of the human liver (33). The localization of thep97 molecule was also shown to be distinct from that of Tf and the TR (33).As well, p97 was found to be present on central vein walls in fetal liver tissue,which was not demonstrated previously (33). These findings are importantbecause they are the first reports of the presence of the p97 molecule on thecell surface of tissues which were previously believed to express none of themolecule. In our own laboratory, it has been shown that the p97 molecule isexpressed by capillary endothelial cells in the normal human brain (Food etal, unpublished observations). As with the liver data, this finding contradictspreviously published data. It would appear, then, that the p97 molecule maybe distributed in normal tissues as well as in all of the cancer tumors and celllines which have been previously described.While p97 does not appear to be as strictly tumor specific as was oncebelieved, there have been several investigations on the possibleimmunotherapeutic properties of the molecule. It was shown that conjugatesof an anti-p97 MAb and the toxin, ricin A-chain, were useful in killinghuman melanoma cells which expressed p97 at more than 80,000 moleculesper cell (34). This effect was not evident in cells which expressed less than5,000 molecules of p97 per cell, indicating that the presence of p97 at lowlevels in normal tissues may not be a factor when considering the efficacy ofimmunotoxins (34). The construction of a recombinant vaccinia virus, whichwas shown to cause infected tissue culture cells to express p97 at a high level,allowed the investigators to attempt to immunize against melanoma (35). Itwas found that the immunization of both mice and monkeys with therecombinant construct resulted in both humoral and cell mediated immune6responses (36). Monkeys are known to express a low level of cross-reactivep97 (36), therefore, it appears that the low level of p97 expression in normaltissues does not interfere with the therapeutic effect being investigated.Another model involving mouse melanoma metastases has also beeninvestigated (24). In this case the generation of a T-cell population reactive top97 was observed, after immunization with the recombinant vaccinia virusconstruct. The result of subsequent adoptive therapy experiments was thatpulmonary melanoma metastases were eradicated in mice treated with theseT-cells (24). There appears, then, to be some potential for the use of p97 as atarget for various therapeutic reagents despite differing interpretations of thenormal distribution status of the molecule.In addition to the Tf-like domain, the deduced sequence of p97 has ahydrophobic segment at its C-terminal which was thought to allow themolecule to be inserted into the plasma membrane (27). Using the Eisenbergalgorithm (37), which predicts membrane associated domains, analysis of theputative proteinacious membrane attachment segment of the p97 moleculehas been completed (Food et al, unpublished observations). This analysissuggests that the putative transmembrane domain of the p97 protein will notact as a membrane anchor, which implies that p97 is anchored to the plasmamembrane by a different mechanism. Though the results of this analysishave never been interpreted in this manner, further investigation revealedthat if a hydrophobic-like transmembrane region is present in a monomericprotein but not predicted to be a hydrophobic transmembrane domain by thisalgorithm, the examined protein is likely attached by a glycosylphosphatidylinositol (GPI) anchor. This analysis has been consistent for anumber of GPI-anchored proteins, including Thy-i, Ly-6 and Qa-2 (Food et al,unpublished observations).7Classically, plasma membrane associated proteins are imbedded in themembrane via a hydrophobic interaction between the lipid bilayer and eitheran alpha helical transmembrane polypeptide segment or a less complexstretch of hydrophobic or uncharged residues (38). Recently, however, manyproteins have been described that are attached to the plasma membrane via aGPI-anchor (38). Proteins with GPI-anchors have been detected in a variety oforganisms ranging from mammals to protozoa, insects, slime molds, andyeast, but not fungi, bacteria and higher plants (38). As well, GPI-anchoredproteins have been described in a variety of cell types, exhibiting considerablediversity in both functional and evolutionary terms (39).The general structural features of GPI-anchors have been summarized(Figure 1). Briefly, the C-terminal amino acid of the protein is linked throughthe x-carboxyl group via an amide linkage to the amino group of thephosphoethanolamine moiety of the GPI-anchor. Thephosphoethanolamine moiety is then linked to a glycan, usually consisting ofmannose and glucosamine, while the glucosamine is linked glycosidically tothe inositol-containing phospholipid. The anchoring of the protein to theplasma membrane is carried out by phosphatidylinositol (38). The completestructure of several GPI-anchored proteins have been determined includingthe variable surface glycoprotein (VSG) from Trypanosoma brucei (40) andmammalian Thy-1(41). It is not surprising that the generalized structure forGPI-anchors is based on work done on these two molecules (38).Many different procedures have been developed for the identificationof GPI-anchored proteins. The most commonly used technique involves theuse of the bacterial enzyme phosphatidylinositol-specific phospholipase C (P1-PLC) for the release of the GPI-anchored protein from the cell surface (38).Another technique involves biosynthetically labelling cells with radiolabelled8Figure 1: General structural features of the GPI-anchor.The general structure of the GPI-anchor with the sites of enzyme actionindicated. This figure has been adapted from the review of Ferguson andWilliams (47). Abbreviations: AA: carboxy-terminal amino acid; EtN:ethanolamine; P: phosphate; glycan: variable carbohydrate portion; G1cNH2:glucosamine. Cleavage sites shown include: HNO2: nitrous aciddeamination; GPI-PLD: site of GPI-PLD activity; PI-PLC/GPI-PLC: site of PTPLC/GPI-PLC activity; PLA2: site of phospholipase A2 activity.GPI-PLDPI-PLC/GPI-PLCNH2Th??9anchor components([3H]-ethanolamine,[3H]-fatty acids,[3H]-inositol, [32P1-phosphate) (38). A further method involving cross reactive determinants onthe released proteins has also been utilized. Generally, the presence of a GPIanchor is proven by a variety of these methods rather than one in particular.The biosynthesis of GPI-anchored proteins has also been described insome detail. It is known that the attachment of the protein to the GPI-anchoroccurs within a minute of completion of polypeptide synthesis (42). Due tothis rapid processing time, it is thought that the protein is attached to apreformed GPI-anchor, in the endoplasmic reticulum by a transamidase (38).The signal which directs the attachment of the GPI-anchor to the protein isthought to exist in the C-terminal region of the protein (38). There is noconsensus sequence for the GPI-anchor attachment, the only features thatoccur regularly are that the C-terminal residues are moderately hydrophobicand that there are a pair of small residues placed 10-12 amino acids to the N-terminal side of this hydrophobic domain (43,44). It is thought that these arethe only requirements for GPI-anchoring, and no other motifs are necessary(44). There have been experiments where the C-terminal regions of GPIanchored proteins are switched with the same regions of standardhydrophobic transmembrane proteins, resulting, after processing, in theattachment of a GPI-anchor to these proteins (45). After the GPI-anchor hasbeen attached to the protein there can also be post attachment events such asthe addition of sugar side chains (38). There can also be defects in the GPIanchor machinery which can, in some cases, result in the release of solubleforms of the protein. In fact, 8 different complementation groups with defectsin the production of the Thy-i molecule have been produced byimmunoselection (46). Since the defect in only one of these classes is in thestructural gene of Thy-i, the biochemical defects in each of the remaining10classes should constitute different elements of the GPI-anchor biosyntheticpathway.One of the most interesting features of GPI-anchored proteins is thatthey can be released from the cell surface by the action of specificphopholipases. Several of these enzymes have been characterized, with themost common the bacterial PT-PLC. These enzymes cleave between theglycerol backbone and the phosphate group of the GPI-anchor (Figure 1) (47).There are also eukaryotic GPI-PLC which are membrane bound and cleaveonly GPI and not PT as is the case for the bacterial P1-PLC (38). These enzymescleave the GPI-anchor at the same place as the bacterial PT-PLC (Figure 1). Thefinal group is the eukaryotic GPT-PLD type of enzymes. Once again theseenzymes are specific to the GPI-anchor only (38), cleaving between theinositol and the phosphate (Figure 1). While this enzyme has been found inthe mammalian plasma, its activity on proteins from intact cell membranes isnot known (48).The fungal metabolite Brefeldin A (BFA) is known to have manydifferent effects on the secretory pathway of mammalian cells (49). The effectsof BFA are still a matter of some debate, but they can be briefly summarized inthis manner: BFA affects the early secretory pathway, resulting in theinhibition of protein transport and the disassembly of the Golgi apparatus(50). It is the loss of the coat protein 3-cop from Golgi membranes in thepresence of BFA, followed by formation of tubular connections between theGolgi cisternae which results in movement of Golgi enzymes to the ER (49).It is also known that BFA affects the trans-Golgi network and the endosomalsystem (49). While BFA is known to inhibit the transport of membrane andsecreted proteins to their appropriate destinations the effect on GPI-anchoredproteins is not well documented.11In this work, the structure which is responsible for the attachment ofthe p97 molecule to the surface of melanoma cells was investigated.12MATERIALS AND METHODS:Cell culture and Flow Cytometry:The human melanoma line SK-MEL-28 (HTB 72) and the mouselymphoma line EL-4 (TIB 39) were obtained from the American Type CultureCollection (ATCC). These lines were maintained in Dulbecco’s ModifiedEagle Medium (DMEM) supplemented with 10% fetal bovine serum (FBS), 20mM HEPES, 100 U/ml penicillin, 100 .tg/ml streptomycin, 2 mM Lglutamine, and 50 p.M 2-mercaptoethanol. The Chinese Hamster Ovary(CHO) cell lines WTB and TRVB were obtained from Dr. F. Maxfield of NewYork University, New York City. These lines were maintained in Hams F12medium supplemented with 10% FBS, 20 mM HEPES, 100 U/ml penicillin,100 p.g/ml streptomycin, and 2 mM L-glutamine. The transfected WTB andTRVB cell lines expressing human p97 were grown in the same media withthe addition of G418 sulfate (Gibco) at a concentration of 800 p.g/ml. All celllines were incubated at 37°C in a 5% C02 humidified atmosphere. Whennecessary, adherent lines were released by treatment with versene.L235 is a hybridoma cell line (HB 8446) from ATCC secreting a mouseMAb (IgG) that reacts with the human p97 molecule. OKT9 is a hybridomacell line (CRL 8021) from ATCC that produces a mouse MAb (IgG) that reactswith the human TR. The T24/3i.7 cell line was obtained from Dr. R. Hyman,of The Salk Institute, in San Diego. This hybridoma secretes a rat MAb (IgG)which reacts with mouse Thy-i. The yEl /9.9.3 cell line was obtained from Dr.F. Takei of The University of British Columbia, in Vancouver. Thishybridoma secretes a rat MAb (IgG) which reacts with the mouse TR. Inmost cases tissue culture supernatants were used as the source of antibody.The reactivity of the various MAbs to the human and mouse antigensof interest in this study were investigated. In these experiments the cells were13counted (106 cells/tube) and washed twice in fluorescence activated cellsorting (FACS) buffer, which consisted of DMEM containing 0.5% (wt/vol)bovine serum albumin, 20 mM HEPES, and 20 mM NaN3. The cells wereincubated with the various MAbs for 45 mm at 4°C, then washed and labelledwith the appropriate fluoresceinated secondary antibody for 45 mm at 4°C.The cells were then washed and fixed in 1.5% (vol/vol) p-formaldehyde inPBS. A Becton-Dickinson FACScan flow cytometer was used to measure 5000events per sample. The conversion of log scale to linear scale values wasaccomplished by using the formula: linear mean fluorescence = 10 (log meanfluorescence/256 channels) (51). The fluorescence intensities werenormalized with respect to unstained control samples and presented as meanlinear fluorescences.For some of the procedures in this work it was necessary to have asource of purified MAb. The tissue culture supernatants for MAb L235, C andOKT9 were centrifuged and filtered. The MAbs were then purified with aProtein G Sepharose column (MAbTrap G/Pharmacia) by the procedurerecommended by the manufacturer. The purified MAb were then tested: forpurity by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDSPAGE) under denaturing conditions, for protein concentration byspectrophotometry, and for activity by FACS.Sensitivity of p97 to release by bacterial enzymes as shown by Flow Cytometry:The purified P1-PLC used in this work was the kind gift of Dr. M. G.Low, of Columbia University in New York City. This enzyme was isolatedfrom culture supernatants of Bacillus subtilis (BG2320), transfected with thePT-PLC gene from Bacillus thuringiensis, purified according to Low (52). Atitration of purified bacterial PT-PLC activity on p97’ on SK-MEL-28 cells rangedfrom 64% release (0.17 U/b6 cells), to 88% release (0.85 U/b6 cells), and 95%14release (1.7 U/b6 cells). The purified P1-PLC was used at a concentration of1.7 U/b6 cells (17 U/ml) in FACS buffer. The crude PT-PLC was theunpurified culture supernatant from the Bacillus subtilis (BG2320)transfected with the PT-PLC gene from Bacillus thuringiensis. In theexperiment 100 tl of the culture supernatant was used for 106 cells. Thepronase used in this study was the Type XIV Protease from Streptomycesgriseus (Sigma) at a concentration of 1 mg/ml in FACS buffer. Inexperiments involving these enzymes the cells were counted, washed twicein FACS buffer and then incubated with the enzyme preparation for lh at37°C. The cells were then labelled for, and analyzed by, FACS, as described inthe flow cytometry section.Phase partitioning in Triton X-114:To investigate p97 partitioning in Triton X-1 14, the cell surface proteinsof 8.0 x 106 cells were labelled with 0.4 mg biotinamidocaproate Nhydroxysuccinimide ester (biotin, Sigma) as described (53). The cells werewashed several times in DMEM, divided into two samples that wereincubated for 60 mm at 6°C, in the presence or absence of P1-PLC (1.7 U/106cells) respectively. Both the cell supernatant and the cell pellet weresubsequently processed. The cells were washed once more and lysed in abuffer containing 10 mM Tris-HC1 pH 7.4, 150 mM NaCl, 1% Triton X-114, 1mM phenylmethylsulfonylfluoride (PMSF) and 100 jig/ml lysine to blockbiotin. Triton X-114 (Sigma) was precondensated as described (54). The samebuffer was added to the supernatant. The samples were centrifuged at 12,000 gfor 10 mm at 4°C to remove the cell nuclei and cell debris. The phaseseparation was obtained by incubation at 30°C followed by a centrifugation at3000 g for 3 mm at room temperature. The samples were re-extracted 3 timesin order to improve the separation and the corresponding phases were15pooled. The samples were precleared for 2 h with washed protein A-agaroseand subsequently divided into two halves for immunoprecipitation of p97and TR using L235 and OKT9 MAb respectively, followed by protein Aagarose precoated with rabbit anti-mouse IgG (Jackson ImmunoResearch).After immunoprecipitation, the samples were washed 6 times in 50 mM TrisHC1 pH 6.5, 150 mM NaC1, 2 mM EDTA, and 0.5% NP-40. The proteins wereeluted from beads in SDS-PAGE loading buffer and separated on an 8% SDSPAGE gel under reducing conditions. The proteins were transferred ontoImmobilon membranes (Millipore) by electroblotting, and detected usingperoxidase-conjugated streptavidin (Jackson ImmunoResearch) and thechemiluminescence ECL Western blotting detection system (Amersham)using the conditions recommended by the manufacturers.Construction of transfectant cell lines expressing human p97:The bacterial neomycin resistance gene, which confers resistance toG418 (Gibco) in mammalian cells, was contained in the pWJ218 construct(Figure 2A) from Dr. W. A. Jefferies and Dr. S. Kvist (unpublished). Thehuman p97 expression vector pSV2p97a (Figure 2B), containing the entirecoding region of p97 cDNA driven by the SV4O early promoter, was obtainedfrom Dr. G. Plowman and Dr. K. E. Hellström, of Bristol-Meyers Squibb inSeattle (55). These plasmids were cotransfected into the WTB and TRVB linesby the Lipofectin method (Gibco) following the procedure recommended bythe manufacturer. Using the appropriate antibodies, cell populationsexpressing p97 were analyzed by flow cytometry. These populations werefurther sorted for cells that expressed higher levels of p97 and then were subcloned by limiting dilution. The resultant cell lines were analyzed by FACS toensure high expression levels of p97. The sensitivity of the p97 expressed bythese lines to release by PT-PLC was then determined by FACS.16Figure 2: Summary of cDNA constructs used to transfect CHO lines.A: pWJ218: This plasmid contains the neo-phosphotransferase gene located inthe Tn 5 element, which confers resistance to the antibiotic G418 (geneticin),driven by the SV-40 early promoter.B: pSV2p97a: This plasmid contains the entire coding region of the p97 gene,driven by the SV-40 early promoter/enhancer.APvuB3 kb 4 kb 5.26 kbI I IH I/1 I—.—-.7.0 kb7.56kb 0kb 1kb 2kbI I Ip i s.1....d1 0p4.4soPvuI17Cell surface biotinylation, P1-PLC treatment, and immunoprecipitation:Surface proteins of 3.0 x 106 cells were labelled with 0.2 mg biotin(Sigma) as described (53). The cells were washed several times in DMEM,divided into two samples that were incubated for 60 mm at 6°C in thepresence or absence of P1-PLC (1.7 U/lU6 cells) respectively. Both the cellsupernatant and the cell pellet were subsequently processed. The cells werewashed once more and lysed in 50 mM Tris-HC1 pH 7.4, 150 mM NaC1, 2 mMEDTA, 0.5% NP-40, 1 mM PMSF and 100 p.g/ml lysine to block the excess offree biotin. The same buffer was added to the supernatant. The samples werecentrifuged at 12,000 g for 10 mm at 4°C to remove the cell nuclei and celldebris. The samples were precleared for 2 h with washed protein A-agarose.The p97 was immunoprecipitated with MAb L235 followed by protein Aagarose precoated with rabbit anti-mouse IgG (Jackson ImmunoResearch).After immunoprecipitation, the beads were washed 6 times in 50 mM TrisHC1 pH 6.5, 150 mM NaC1, 2 mM EDTA, and 0.5% NP-40. The proteins wereeluted from the beads in SDS-PAGE loading buffer and separated on an 8%SDS-PAGE gel under reducing conditions. The proteins were transferredonto Immobilon membranes (Millipore) by electroblotting, and detectedusing peroxidase-conjugated streptavidin (Jackson ImmunoResearch) and thechemiluminescence ECL Western blotting detection system (Amersham)using the conditions recommended by the manufacturer.Biosynthetic labelling with[3H]-ethanolamine:The cell line monolayers were biosynthetically labelled for 24 h with[3 H] ethan-1 -ol-2-amine hydrochloride (20 !.tCi / ml, 30.4 Ci / mmol,Amersham) in DMEM containing 5% dialyzed FBS, 20 mM HEPES, 100 U/mlpenicillin, 100 .tg/ml streptomycin, 2 mM L-glutamine, and 50 tM 2-mercaptoethanol. The cells were washed in PBS and lysed in 50 mM Tris-HC118pH 7.2, 150 mM NaC1, 2 mM EDTA and 0.5% NP-40 with 40 p.g/ml PMSF.The lysates were then cleared by centrifugation prior to theimmunoprecipitation. The primary antibodies used were the L235 for p97and the OKT9 for the human TR. Protein A-agarose (BioRad) coated withrabbit anti-mouse IgG antibody (Jackson ImmunoResearch) was added to thesamples and incubated for 8 h at 4°C. The resulting complex was washed in50 mM Tris-HC1 pH 6.5, 150 mM NaC1, 2 mM EDTA, and 0.5% NP-40 andresuspended into SDS-PAGE loading buffer. The samples wereelectrophoresed under reducing conditions on a 10-15% gradient SDS-PAGEgel. After fixation the gel was treated with Amplify (Amersham), dried, andautoradiographed.Effect of BFA treatment on the transport of p97 to the cell surface:The fungal metabolite BFA was obtained from Dr. F. Tufaro, of theUniversity of British Columbia, in Vancouver. The experiment involvedincubating the SK-MEL-28 cells with bacterial P1-PLC (1.7 U/b6cells) for 1 h at37°C and then washing and incubating the cells at 37°C with 5.0 g/ml BFA inthe regular SK-MEL-28 tissue culture media. The cells were removed after 0,20 and 40 h, and then were stained with the appropriate primary andfluoresceinated secondary antibodies, and then were washed, fixed andanalyzed by FACS for cell surface p97 expression.Expression of p97 on human tumor lines in a P1-PLC sensitive form:The human duodenal adenocarcinoma line HuTu-80 (HTB-40) and thehuman colonic adenocarcinoma line CaCo-2 (HTB-37) were also obtainedfrom the ATCC. These lines were maintained in minimum essential media(MEM) with Earl’s balanced salt solution supplemented with 20% FBS, 20mM HEPES, 100 U/ml penicillin, 100 pg/ml streptomycin, 2 mM Lglutamine, and 0.1 mM non-essential amino acids. The cell lines were19screened by a panel of MAb and the results were visualized by FACS. Thesensitivity of the p97 to release by PT-PLC was then determined by FACS.20RESULTS:Antibodies and Enzymes:The MAbs used in this work are summarized in Table I. The reactivityof these MAbs with various human and mouse surface antigens arepresented in Figures 3 to 5. By comparing Figure 3 (SK-MEL-28) with Figure 4(p97aTRVBc3), it was apparent that there were four different MAb availablewhich reacted with p97. The MAbs 96.5 and 133.2 were not available for largescale use, thus only the L235 and C MAb were considered for this study. It isinteresting to note that the MAb A was extremely reactive to the SK-MEL-28cells, but not to the p97aTRVBc3 cells, indicating that it was specific for amelanoma associated antigen that was distinct from p97. In Figure 5 it isshown that the anti-mouse MAbs are specific for the mouse antigens and thatthe anti-human MAbs show no cross reactivity with the mouse antigens.The reverse is the case in Figure 3, where the anti-human MAbs are specificfor the human antigens, while the anti-mouse MAbs show no cross reactivitywith the human antigens.The purified MAbs were titrated against the culture supernatants inFigures 6 to 8. In Figure 6 the purified L235 MAb is shown to react reasonablywell to p97. The same is not the case for the C MAb, where it appeared thatthe activity of the purified MAb is somehow impaired by the purificationprocedure (Figure 7). For this reason the L235 MAb was chosen to be the antip97 MAb for this study. The titration of the anti-TR MAb, OKT9 provided aninteresting result (Figure 8). The use of serum free tissue culture mediainstead of serum containing media did not impair the ability of thehybridoma cells to produce active MAbs. In the future this media will beused to produce MAbs in our laboratory.TableI:SummaryofthepropertiesoftheMAbsusedinthisstudy.UnlessotherwisestatedtheMAbculturesupernatantswereusedat100.tlfor106cells.AntigenNameSpeciesTypeSourceSupplyHumanHLAA2/28PA2.1MouseIgGATCCSupernatantHumanFNRFNRRabbitAntiseraTeliosPurifiedHumanMelanomaAMouseIgGLiaoSupernatantHumanp9796.5MouseIgG2aHellströmPurifiedHumanp97133.2MouseIgG2aHellströmPurifiedHumanp9?L235MouseIgGiATCCBothHumanp97CMouseIgGLiaoBothHumanTROKT9MouseIgGIATCCBothMouseH-2b20-8-4sMouselgG2aATCCSupernatantMouseThy-IT24/37.IratIgGIHymanSupematantMouseTRE1/9.9.3ratIgGITakeiSupernatant22Figure 3: Reactivity of various MAbs with surface antigens of SK-MEL-28cells.The cell surface antigens were labelled with the primary and the appropriatefluoresceinated secondary antibody. After washing and fixing, the sampleswere analyzed by FACS. The results were converted to linear scale andnormalized with respect to unstained negative control samples.SK-MEL-2890807060Mean Linear 50Fluorescence 40302010096.5 133.2 A__IB C L235 OKT9 PA2.laEl/9. T24/39.3 7.1Antibody23Figure 4: Reactivity of various MAbs with surface antigens of p97aTRVBc3cells.The cell surface antigens were labelled with the primary and the appropriatefluoresceinated secondary antibody. After washing and fixing, the sampleswere analyzed by FACS. The results were converted to linear scale andnormalized with respect to unstained negative control samples.p97aTRVBC3700600500Mean Linear 400Fluorescence 3002001000B C L235 OKT9 PA2.1E1/9. T24/39.3 7.1Antibody96.5 133.2 A24Figure 5: Reactivity of various MAbs with surface antigens of EL-4 cells.The cell surface antigens were labelled with the primary and the appropriatefluoresceinated secondary antibody. After washing and fixing, the sampleswere analyzed by FACS. The results were converted to linear scale andnormalized with respect to unstained negative control samples.EL-4E1 /9.9.3AntibodyMean Linearfluorescence300250200150100500OKT9 PA2.1 T24/37.125Figure 6: Titration of purified L235 MAb reactivity against SK-MEL-28 cells.Tissue culture supernatants of the L235 MAb were centrifuged and filtered.The MAb was then purified with a protein G column (Pharmacia). The MAbwas then diluted in FACS buffer and incubated with SK-MEL-28 cells. Theappropriate fluoresceinated secondary antibody was then used. After washingand fixing, the samples were analyzed by FACS. The results were convertedto linear scale and normalized with respect to unstained negative controlsamples. L23512010080Mean Linear 60Fluorescence40200IiiIIII I I ISupernatant 1:10 1:50Antibody Dilution1:100 1:200H26Figure 7: Titration of purified C MAb reactivity against SK-MEL-28 cells.Tissue culture supernatants of the C MAb were centrifuged and filtered. TheMAb was then purified with a protein G column (Pharmacia). The MAb wasthen diluted in FACS buffer and incubated with SK-MEL-28 cells. Theappropriate fluoresceinated secondary antibody was then used. After washingand fixing, the samples were analyzed by FACS. The results were convertedto linear scale and normalized with respect to unstained negative controlsamples. c605040Mean Linear 30Fluorescence20100Supernatant 1:10 1:50 1:100 1:200Antibody Dilution27Figure 8: Titration of purified OKT9 MAb reactivity against HuTu-80 cells.Tissue culture supernatants of the OKT9 MAb were centrifuged and filtered.The MAb was then purified with a protein G column (Pharmacia). The MAbwas then diluted in FACS buffer and incubated with HuTu-80 cells. Theappropriate fluoresceinated secondary antibody was then used. After washingand fixing, the samples were analyzed by FACS. The results were convertedto linear scale and normalized with respect to unstained negative controlsamples.OKT918161412Mean Linear 10Fluorescence 86420Supemat 1:02 1:10 1:50antAntibody Dilution• NormalEl Serum Free1:10028The effects of the various PT-PLC preparations and pronase on thehuman and mouse cell surface antigens of interest were investigated (TableII). While the crude and pure PT-PLC produced similar results, it was clearthat the pure P1-PLC did not affect the TR as much as the crude PT-PLC did.This was perhaps due to contamination of the crude P1-PLC preparation withnon-specific proteases from the recombinant Bacillus subtilis which was usedto produce the P1-PLC. The purified PT-PLC preparation was selected for usein this work, due to the apparent lack of contamination with proteases, and isreferred to as simply P1-PLC in subsequent text, Tables and Figures.Release of p97 by bacterial P1-PLC:Studies have shown that bacterial PT-PLC can cleave GPI-anchors andrelease the proteins in a soluble form from intact cells. This cleavage is themost common criterion used in GPI-anchor identification (38). Thesensitivity of p97 to treatment with purified bacterial PT-PLC was tested bystaining SK-MEL-28 cells with the L235 MAb and measuring the surfaceexpression of p97 by flow cytometry. The release of p97 by bacterial PT-PLCtreatment is represented by the profiles in Figure 9. The results wereconverted from logarithmic to linear scale and are expressed as a percentageof the control (Table ITI). These values indicate that the amount of p97expression on SK-MEL-28 cells was decreased to 10% of initial levels bytreatment with bacterial P1-PLC. In contrast, the expression of the human TRon SK-MEL-28 cells was not at all changed by treatment with bacterial PT-PLC.An example of a protein known to be GPI-anchored is Thy-I, the mouselymphocyte antigen (56). The release of Thy-i by bacterial PT-PLC treatment isalso represented by the profiles in Figure 9, and summarized in Table III. Theexpression of Thy-i on the EL-4 cells was reduced to 30% of initial levels by29Table II: Effect of various PT-PLC preparations and pronase on human andmouse cell surface antigens.Cells were incubated for I h at 37°C in the appropriate enzyme preparation.The human p97 was labelled with the L235 MAb, the human TR was labelledwith the OKT9 MAb, and the human fibronectin receptor was labelled withthe antisera from Telios. The mouse Thy-i was labelled with the T24/37.iMAb, the mouse TR was labelled wih the yEi/9.9.3 MAb, and the mouse H2bwas labelled with the 20-8-4s MAb. The appropriate fluoresceinated secondaryantibodies were then used. After washing and fixing, the samples wereanalyzed by FACS. The results were converted to linear scale and normalizedwith respect to unstained negative control samples and the values expressedas percentages of untreated positive control samples.Enzyme Fluorescence Intensity (% of control)Treatment Human Mousep97 TR FNR Thy-I TR H-2bCrude P1-PLC 22 89.5 101 22 77 93Pure P1-PLC 2 90 101 21 98 95Pronase 72 1 190 115 1 9230Figure 9: Effect of bacterial PT-PLC on cell surface antigens.SK-MEL-28 cells were treated with (c,f), and without (b,e), bacterial PT-PLC.The cells were then labelled for human p97 with the L235 MAb (b,c), and forhuman TR with the OKT9 MAb (e,f). The negative controls (a,d) were no firstantibody stainings of SK-MEL-28 cells which were not treated with bacterialPT-PLC. EL-4 cells were treated with (i,1), and without (h,k), bacterial PT-PLC.The cells were then labelled for mouse Thy-i with the T24/37.i MAb (h,i),and for the mouse TR with the yEl /9.9.3 MAb (k,l). The negative controls(g,j) were no first antibody stainings of EL-4 cells which were not treated withbacterial P1-PLC. After incubations with the appropriate fluoresceinatedsecondary antibodies, the cells were washed, fixed and analyzed by FACS. Thelog scale profiles are presented in this figure.SK-MEL-28EL-4ControINoP1-PLCTreatmentPI-PLCTreatmentp97TRThy-iLogofFluorescenceIntensityTRa) S. C)I1.1______H‘k.••..HA32Table III: Effect of bacterial PT-PLC and pronase on human and mouse cellsurface antigens.Cells were incubated for 1 h at 37°C in FACS buffer and 17 U/mi P1-PLC or 1mg/mi pronase. The human p97 was labelled with the L235 MAb and thehuman TR was labelled with the OKT9 MAb. The mouse Thy-i was labelledwith the T24/37.i MAb while the mouse TR was labelled with the yEl /9.9.3MAb. The appropriate fluoresceinated secondary antibodies were then used.After washing and fixing, the samples were analyzed by FACS. The resultswere converted to linear scale and normalized with respect to unstainednegative control samples and the values expressed as percentages (±s.d.) ofuntreated positive control samples. The data presented here are the result offive independent experiments.Antigen Cells Treatment Fluorescence Intensity(% of control)p97 SK-MEL-28 P1-PLC 10.8±2.6p97 SK-MEL-28 Pronase 82.6±17.7TR SK-MEL-28 P1-PLC 120.5±16.2TR SK-MEL-28 Pronase 15.9±10.8Thy-I EL-4 P1-PLC 32.5±6.4Thy-i EL-4 Pronase 136.4±20.2TR EL4 P1-PLC 93.1±9.4TR EL4 Pronase 1.9±1.033treatment with the bacterial PT-PLC, while the expression of the mouse TR onthe EL-4 cells was not changed by the bacterial PT-PLC treatment.To further support the conclusion that the bacterial PT-PLC preparationdoes not contain a non-specific protease activity, pronase was used to treatcells which were then stained for either the human (p97 or TR) or the mouse(Thy-i or TR) antigens. The human TR was sensitive to the effects of thisenzyme, while p97 was relatively insensitive (Table ITT). The Thy-i moleculewas P1-PLC sensitive and pronase resistant on the mouse lymphoma line EL-4, while the mouse TR was pronase sensitive and P1-PLC resistant (Table ITT).The results of these control experiments are in agreement with publisheddata (56).Partitioning in Triton X-114:The technique of phase separation in Triton X-114 can be used to assessthe amphipathic or hydrophilic character of a protein and is especially usefulto identify GPI-anchored proteins (38). This technique is based on the abilityof the detergent Triton X-114 to partition into two phases: a detergent richphase and a detergent poor phase. Amphipathic proteins which possess ahydrophobic membrane anchor primarily partition into the detergent richphase, whereas hydrophilic proteins partition into the aqueous phase.Proteins with intact GPI-anchors will also partition into the detergent richphase.In order to investigate p97 partitioning in Triton X-114, the cell surfaceproteins were labelled with biotin, washed, and incubated in the presence orabsence of bacterial P1-PLC. Both the cell supernatant and the cell pellet wereprocessed for phase separation. The p97 and TR were immunoprecipitated byMAb L235 and OKT9, respectively, followed by Protein A-agarose. Theproteins were transferred onto a membrane by electroblotting and detected34Figure 10: Phase separation of p97 and TR in Triton X-114 solution.SK-MEL-28 cell surface proteins were labelled with biotin. The cells weresubsequently washed in DMEM and incubated in the presence (+) or absence(—) of P1-PLC (1.7 U/lO6cel s) for 1 h at 6°C. Proteins from the cell pellet (P) orthe cell supernatant (S) were separated in Triton X-114 solution, and p97 andTR were immunoprecipitated from both the aqueous phase (A) or thedetergent phase (D), as described in the Materials and Methods.p97 TRP1-PLCkDa97.4 —68.0 —43.7 —35using peroxidase conjugated streptavidin and the chemiluminescence ECLWestern blotting detection system. Figure 10 shows that all p97 moleculesexpressed at the surface of untreated human melanoma SK-MEL-28 cellspartition into the detergent-rich phase. No p97 was detected in thesupernatants of untreated cells. Treatment with bacterial P1-PLC led to thepartitioning of p97 into the aqueous phase of the cell supernatant sample,indicating that the protein was cleaved from the plasma membrane andreleased as a hydrophilic form. No p97 could be detected in the bacterial PT-PLC treated cell pellet, indicating that most molecules were bacterial PT-PLCsensitive and that p97 is not simultaneously expressed in a transmembraneand GPI-anchored form at the cell surface. In contrast to p97’, the TR, which isinserted in the membrane through a hydrophobic peptide segment, is notaffected by bacterial PT-PLC. The amphiphilic structure causes the protein topartition in both phases after separation.Construction of transfectant cell lines expressing human p97 and sensitivityof p97 to release by bacterial P1-PLC:In order to examine whether the processing signals for GPI-anchorattachment reside in the p97 sequence, and to demonstrate clearly thespecificity of the L235 MAb used in this study, the p97 cDNA was transfectedinto the CHO lines WTB and TRVB. After sub-cloning by limiting dilution,four lines (p97aWTBc3, p97aWTBc7, p97aTRVBc3, p97aTRVBc6) which stablyexpress p97 were isolated. The development of these cell lines from theparental CHO lines is summarized by the FACS profiles in Figures 11 and 12.The treatment of these lines expressing human p97 with bacterial PT-PLCresulted in a decrease in p97 expression at the cell surface (Figure 13). Theexpression of p97 decreased to less than 20% of untreated levels in all four ofthe transfectant cell lines (Table IV).Figure11:SummaryofthedevelopmentoftheCHOlineWTBexpressinghumanp97.WTBcellswerestainedforp97bytheL235MAb(b,d,f,h)andthen,afterincubationwiththeappropriatefluoresceinatedsecondaryantibody,thecellswerewashed, fixedandanalyzedbyFACS.Thenegativecontrols(a,c,e,g)werenofirstantibodystainingsofWTBcells.Thelogscaleprofilesarepresentedinthisfigure.WTB________________________a)-o E________________________________________________________D________________________________________________________z a) 0Controlp97BulkSortedSubclonedC’c,g;I.—I.IH,.bIfhII..I.IL1HLI.1I9—,ILogofFluorescenceIntensityFigure12: SummaryofthedevelopmentoftheCHOlineTRVBexpressinghumanp97.TRVBcellswerestainedforp97bytheL235MAb(b,d,f,h)andthen,afterincubationwiththeappropriatefluoresceinatedsecondaryantibody,thecellswerewashed, fixedandanalyzedbyFACS.Thenegativecontrols(a,c,e,g)werenofirstantibodystainingsofTRVBcells.Thelogscaleprofilesarepresentedinthisfigure.TRVBa) 2 D z ci) 0Controlp97BulkSortedSubclonedcgIIII.I,,o“.9Idfi.1I.,.,.LogofFluorescenceIntensityFigure13:Effectofbacterial PT-PLConhumanp97expressedonthecellsurfaceofvariousCHOlines.ThetransfectantCHOcellsweretreatedwith(c,f,i,l),andwithout(b,e,h,k),bacterialPT-PLC.Thecellswerethenlabelledforhumanp97withtheL235MAb(b,c,e,f,h,i,k,l).Afterincubationwiththeappropriatefluoresceinatedsecondaryantibody,thecellswerewashed, fixedandanalyzedbyFACS.Thenegativecontrols(a,d,g,j)werenofirstantibodystainingsofthetransfectantCHOcellswhichwerenottreatedwithbacterialPT-PLC.Thelogscaleprofilesarepresentedinthisfigure.p97aWTBC3p97aWTBc7p97aTRVBc3p97aTRVBc6-dgaControlNoP1-PLCTreatmentP1-PLCTreatmentAci-Q 2 D 0hA ........ /cIILogofFluorescencelntensty39Table IV: Effect of bacterial PT-PLC on human p97 expressed on the cell surfaceof various CHO lines.Cells were incubated for I h at 37°C in FACS buffer and 17 U/ml PT-PLC. Thehuman p97 was labelled with the L235 MAb. The appropriate fluoresceinatedsecondary antibody was then used. After washing and fixing, the sampleswere analyzed by FACS. The results were converted to linear scale andnormalized with respect to unstained negative control samples and thevalues expressed as percentages (±s.d.) of untreated positive control samples.The data presented here are the result of three independent experiments.Antigen Cells Treatment Fluorescence Intensity(% of control)p97 p97aWTBc3 P1-PLC 6.8±4.0p97 p97aWTBc7 P1-PLC 7.5±5.1p97 p97aTRVBc3 P1-PLC 13.9±4.3p97 p97aTRVBc6 P1-PLC 19.4±10.340Figure 14: Effect of bacterial P1-PLC on human p97 expressed on the cellsurface of SK-MEL-28 and p97aWTBc7 lines.Cell surface proteins were labelled with biotin and the cells incubated in thepresence (÷) or absence (—) of bacterial P1-PLC (1.7 U/lU6 cells) for 1 h at 6°C.Subsequently, p97 was immunoprecipitated from the cell pellets (Panel A) orfrom the cell supernatants (Panel B) and detected as described in the Materialsand Methods. The cell lines analyzed were 1: WTB, 2: p97aWTBc7, 3: 5K-MEL-28.AkDa PI.PLC200.0 —97.4 —68.0 —1 2 3- 1 -1 -IBI-I4.3.7 —41The effect of bacterial PT-PLC and the specificity of the L235 MAb wasfurther characterized in Figure 14 where surface proteins from either WTB,p97aWTBc7 or SK-MEL-28 cells were labelled with biotin and p97 wasimmunoprecipitated and analyzed on SDS-PAGE under reducing conditions.The proteins were transferred onto a membrane by electroblotting anddetected using peroxidase conjugated streptavidin and thechemiluminescence ECL Western blotting detection system. A single proteinof 95-97,000 daltons molecular mass was immunoprecipitated from the cellline transfected with the p97 cDNA that was absent in the parentaluntransfected cell line WTB. The same protein was detected in the humanmelanoma cell line SK-MEL-28. Treatment of both SK-MEL-28 or thetransfected p97aWTBc7 cells with bacterial PT-PLC resulted in a large loss ofthe protein from the cell surface (Figure 14A). The released protein could berecovered in the cell supernatant (Figure 14B). Under the conditions used inthis experiment, no difference in the molecular mass between the plasmamembrane associated form and the released form could be detected.Biosynthetic labelling:It can not be excluded from the FACS and the biotinylationexperiments that the decrease in the expression observed after bacterial PT-PLCtreatment is an indirect effect due to the association of p97 with another PT-PLC sensitive protein at the cell surface. To address this issue, the cells inculture were biosynthetically labelled with[3H]-ethanolamine, which isknown to be a component of the phospholipid moiety of GPI-anchoredproteins (38). The cells were lysed, p97 or TR were immunoprecipitated, andrun on SDS-PAGE under reducing conditions. The p97 and TR moleculeswere immunoprecipitated by the L235 and OKT9 MAbs respectively, in theSK-MEL-28 cell line (Figure 15). The p97 is labelled by the[3H1-ethanolamine42Figure 15: Biosynthetic labelling of p97 with[3H1-ethanolamine in SK-MEL-28cells.SK-MEL-28 cell monolayers were biosynthetically labelled in duplicate with[3H]-ethanolamine and proteins were immunoprecipitated with theappropriate MAb. These complexes were separated on a 10-15% SDS-PAGEgel and autoradiographed. In Lanes 1 and 2 p97 was immunoprecipitatedwith the L235 MAb. The observed difference in labelling intensity may be dueto slight differences between the individual plates used in theimmunoprecipitation. Lanes 3 and 4 show the immunoprecipitation of thehuman TR with the OKT9 MAb.kDa 1 2 3 4200.097.469.046.030.043(lanes 1+2) while theTR is not (lanes 3+4). The p97 in both the p97aWTBc3line (Figure 16: lane 3) and the p97aTRVBc3 (Figure 17: lane 3) line was alsolabelled by the[H1-ethanolamine.Effect of BFA on the transport of p97 to the cell surface:The effect of the fungal metabolite BFA on the transport of p97 to thecell surface was also investigated. The SK-MEL-28 cells were treated with PT-PLC (1.7 U/106 cells) in DMEM and then were allowed to recover in thepresence or absence of BFA. It is clear that the 5 .tg/ml BFA concentrationdisrupted the transport of the p97 molecule to the cell surface (Table V). Theresults of these experiments with regards to the blocking of surface expressionof a GPT-anchored protein by BFA are in agreement with previous work onalkaline phosphatase (57).Expression of p97 on human tumor cell lines in a P1-PLC sensitive form:The presence of p97 on the cell surface of two different humanintestinal tumor lines was investigated using a variety of MAb against p97.The results of these experiments were consistent for both lines in that p97 wasexpressed at a fairly low level compared to the human TR (Figures 18,19).The effects of bacterial P1-PLC treatment on these antigens are summarized inTable VI. It is again evident that p97 is sensitive to treatment with P1-PLC,while the TR is not sensitive to the effects of this enzyme.44Figure 16: Biosynthetic labelling of p97 with[3H]-ethanolamine inp97aWTBc3 cells.WTB (Lane 1,2) and p97aWTBc3 (Lane 3,4) cell monolayers werebiosynthetically labelled with[H]-ethanolamine and proteins wereimmunoprecipitated with the appropriate MAb. The complex was thenanalyzed as in Figure 14. Lane 3 shows the immunoprecipitation of p97 withL235 MAb. Lanes 1 (L235), 2 (OKT9), and 4 (OKT9) indicate that the MAb usedin this experiment do not cross react with labelled hamster antigens.kDa 1 2 3 4200.097.469.046.030.045Figure 17: Biosynthetic labelling of p97 with[3HJ-ethanolamine inp97aTRVBc3 cells.TRVB (Lane 1,2) and p97aTRVBc3 (Lane 3,4) cell monolayers werebiosynthetically labelled with[H1-ethanolamine and proteins wereimmunoprecipitated with the appropriate MAb. The complex was thenanalyzed as in Figures 14 and 15. Lane 3 shows the immunoprecipitation ofp97 with L235 MAb. Lanes 1 (L235), 2 (OKT9), and 4 (OKT9) indicate that theMAb used in this experiment do not cross react with labelled hamsterantigens.kDa 1 2 3 4200.097.469.046.030.046Table V: Effect of BFA on the transport of p97 to the surface of SK-MEL-28cells.SK-MEL-28 cells which had been treated with 17 U/ml P1-PLC for one h at37°C were then allowed to recover at 37°C with 5.0 .tg/ml BFA in DMEM orDMEM alone. At various times (0, 20, 40 h) cells were removed andincubated with the L235 MAb. After incubation with the appropriatefluoresceinated secondary antibody, the cells were washed, fixed, and analyzedby FACS. The results were converted to linear scale and normalized withrespect to unstained negative control samples, and the values expressed aspercentages of untreated positive control samples.Treatment Treatment Time (h) % of p97 remainingInitial Recovery on cell surfaceNo No 0 100.00No No 20 100.00No No 40 100.00P1-PLC No 0 2.30PT-PLC No 20 8.81PT-PLC No 40 52.80PT-PLC BFA 0 2.30PT-PLC BFA 20 0.30P1-PLC BFA 40 0.00No BFA 0 100.00No BFA 20 49.83No BFA 40 15.0047Figure 18: Reactivity of various MAbs with surface antigens of CaCo-2 cells.The cell surface antigens were labelled with the primary and the appropriatefluoresceinated secondary antibody. After washing and fixing, the sampleswere analyzed by FACS. The results were converted to linear scale andnormalized with respect to unstained negative control samples.CaCo-2ILPA 2.1 OKT9 L235 CAntibodyMean LinearFluorescence100 -90 -8070 -60 -50 -40 -30 -20 -10 -0-96.5 133.248Figure 19: Reactivity of various MAbs with surface antigens of HuTu-80 cells.The cell surface antigens were labelled with the primary and the appropriatefluoresceinated secondary antibody. After washing and fixing, the sampleswere analyzed by FACS. The results were converted to linear scale andnormalized with respect to unstained negative control samples.HuTu-8090807060Mean Linear 50Fluorescence 403020100C 133.2Antibody49Table VI: Effect of bacterial PT-PLC on human p97 expressed on the cell surfaceof various human intestinal lines.Cells were incubated for 1 h at 37°C in FACS buffer and 17 U/mi PT-PLC. Thehuman p97 was labelled with the L235 MAb and the human TR was labelledwith the OKT9 MAb. The appropriate fluoresceinated secondary antibodieswere then used. After washing and fixing, the samples were analyzed byFACS. The results were converted to linear scale and normalized withrespect to unstained negative control samples and the values expressed aspercentages (±s.d.) of untreated positive control samples. The data presentedhere are the result of three independent experiments, with the exception ofthe HuTu-80 p97 determination which was completed once.Antigen Cells Treatment Fluorescence Intensity(% of control)p97 CaCo-2 P1-PLC 35.0±22.4TR CaCo-2 P1-PLC 94.0±17.9p97 HuTu-80 P1-PLC 15.4TR HuTu-80 P1-PLC 87.0±8.150DISCUSSION:In this work it is shown that p97 is attached to the cell surface via a GPIanchor. The first section demonstrates that p97 is sensitive to the effects ofbacterial PT-PLC. This technique is commonly used to test for the presence ofa GPT-anchor. In this case p97 was clearly shown to be sensitive to the effectsof purified bacterial P1-PLC in the human melanoma line SK-MEL-28 (Figure9, Table III). The effect was not due to contaminating protease activity, as wasshown by the relative insensitivity of p97 to pronase. The human TR wasused as an example of a typical transmembrane protein, which was bacterialP1-PLC resistant and pronase sensitive. The detergent Triton X-1 14 was usedto partition the p97 and TR, a further test for the presence of a GPI-anchor(Figure 10). It is clear from this experiment that p97 possesses a lipid anchor,which permits it to partition in the detergent phase of the pellet after nobacterial P1-PLC treatment, and the aqueous phase of the supernatant afterbacterial P1-PLC treatment.The specificity of MAb L235 for p97’, and the presence of theinformation necessary to confer GPI-anchor attachment in the p97 cDNA, areproven by the transfections of the CHO cells with the p97 cDNA. As well, themachinery responsible for the synthesis of the GPI-anchor, and its attachmentto the p97 protein, is shown to be present in the CHO line, in a form thatrecognizes some signal on the p97 pre-protein. The data for the transfectantcell lines where bacterial P1-PLC release was visualized by FACS (Figure 13,Table IV), by cell surface labelling (Figure 14), and the biosynthetic labellingwith[3H]-ethanolamine (Figures 16,17), all indicate that the L235 MAbrecognized the protein product of the p97 cDNA. It is also clear from thesefigures that the expression of p97 by the transfectant lines is considerablygreater than for the melanoma line. Since the experiments showed that the51same form of the p97 molecule was present on the transfectant lines as on themelanoma line, it is presumed that this form maintains the properties of themolecule. These lines should, therefore, be useful in the study of thefunctional aspects of the p97 molecule.It was also demonstrated by biosynthetic labelling that p97 is not simplyassociated with a GPI-anchored protein at the cell surface (Figures 15-17). Thepresence of[3H1-ethanolamine in the p97 from both melanoma andtransfected CHO lines indicated that p97 is indeed GPI-anchored. Thesignificance of this GPI-anchor with respect to intracellular transport wasinvestigated with BFA (Table V). It is obvious that the lipid anchor does notenable the p97 molecule to procede to the cell surface via a transportmechanism distinct from the standard protein pathway, as is the case forcholesterol and phosphatidylethanolamine (58,59). To complete theinvestigation, the presence of p97 in a bacterial PT-PLC sensitive form on twohuman intestinal carcinoma lines was demonstrated (Figures 18,19 and TableVI).Unlike Tf, which is a soluble protein whose uptake is mediated by theTR, p97 is expressed at the cell surface. The structure of the GPI-anchorconfers important biophysical differences between this and the conventionalmethod of protein anchoring (e.g.: anchor via a protein transmembranehydrophobic region), therefore, the finding that p97 is GPI-anchored hasseveral consequences. A major feature of GPT-anchored proteins whichdistinguishes them from other membrane proteins is their ability to bereleased from membranes by the action of bacterial PT-PLC. The identificationof mammalian GPT-anchor hydrolysing enzymes (48) suggests that onefunction of the anchor might be to allow the rapid release of the protein. TheGPI-anchor could provide a very specific way for a cell to regulate the52expression of p97 at the cell surface at a post-translational level, using theanchor as a quick release method of down regulation at the cell surface, or upregulation of the protein in circulation (39). In fact, the GPI-anchor of p97may be present to allow the cell to release a soluble form of p97 in response toa cue or signal. The existence of soluble p97 in the spent tissue culturemedium of human melanoma cells has been previously documented (29).We have also detected a water soluble form of the p97 protein in tissueculture medium, and the source from which this form of the p97 moleculeoriginates is currently under investigation (data not shown).A further consequence of lipid anchoring is an inherent increase inlateral mobility in the plane of the membrane and exclusion from coated pitsinvolved in receptor mediated endocytosis (38). It has been shown that theGPI-anchored folate receptor is excluded from the clathrin coated pits, butassociates with small invaginations on the cell surface termed caveolae,which are capable of mediating the internalization of its ligand (60-62). It hasalso been shown that a GPI-anchored form of CD4 is internalized by amechanism which is different from that of the folate receptor in that thereceptors are clearly internalized in vesicles (45). It is possible that p97 couldbe directly involved in iron uptake in melanoma cells. While cellular ironuptake has been shown to be mediated mainly by the Tf/TR pathway, there isevidence for a non-Tf mediated pathway of iron incorporation in various celllines (12-15). Since iron is an absolute requirement for cell growth, there areseveral potential consequences of a p97 mediated iron uptake system in termsof the development of melanoma. Melanoma cells may have a growthadvantage over normal cells by combining two iron uptake pathways, onewhich is mediated by TR and the other which is mediated by p97.53Alternatively, p97 could deprive normal cells surrounding the tumor of theiriron supply, resulting in cell death and the further expansion of the tumor.The potential iron uptake capacity of p97 has been investigated. It wassuggested that p97 was responsible for an iron uptake activity evident in SKMEL-28 cells (16). In a subsequent report by the same group (17), it was statedthat p97 was not contributing to iron utilized by the cell. The next reportfrom this group concluded that p97 was not involved in the uptake of ironfrom inorganic complexes (18). The most recent report from this groupinvestigated the effect of ferric ammonium citrate and desferrioxamine onthe uptake of iron by the SK-MEL-28 cells (19). It is suggested that the ironuptake evident is due to the processes described previously (16). This grouphave not yet demonstrated that the non-Tf bound membrane iron uptakeactivity is due to the action of p97. An experiment which might address thisproblem is to examine the activity of this non-Tf bound membrane uptakesystem after treatment with bacterial P1-PLC. Presumably, if p97 wasresponsible for the observed activity, the iron uptake activity would beabolished by the bacterial P1-PLC treatment. One further problem which isapparent in these papers is that the iron uptake component which isattributed to p97 is stated to be pronase sensitive (16). The data presented inthis work indicate that p97 is relatively insensitive to the effects of pronase. Ifboth sets of data are correct, one possible explanation for this situation is thatthe activity which they attribute to p97 is not due to p97, but some other nonTf membrane iron uptake system.The original description of the iron binding capacity of p97 was doneon SK-MEL-28 cells which were incubated with [59Fe] as FeC13, and then celllysates were exposed to various MAb and the mixture was passed over acolumn (25). While it was clear that p97 could bind iron, the number of54atoms of iron bound per molecule of p97 was not known. It was recentlysuggested that only one functional iron-binding site exists on the p97molecule (63). It is thought that the iron binding properties of the N-terminalsite of p97’ are conserved, but those of the C-terminal site are not (63). Sincethe purification scheme used in this experiment appears to isolate only themembrane bound form of p97, perhaps the result is only applicable to theGPI-anchored form of p97 and not the soluble form of p97. An interestingstudy would be to carry out the same iron binding experiment on bacterial PT-PLC released, purified p97. It is possible that the presence of the plasmamembrane so close to the C-terminal site is the only factor preventing theinteraction of the iron with the binding site (64). Alternatively, the p97molecule may only require the N-terminal site for whatever function it isinvolved in (63).One area of concern in regards to the single atom of iron/p97 moleculehypothesis is the wavelength used to measure the amount of iron bound.The electronic absorption spectrum of iron-saturated human serum Tf hasbeen reported to peak at 464 nm (63) or 473 nm (65). Unfortunately, similardata does not exist for the p97 molecule. The single atom of iron/p97molecule determination used the max of 464 nm (63), while another grouphas suggested that the max of 420 nm is more appropriate for p97 (65). Thisgroup worked with site directed mutants of the N-terminal half of human Tf.In one of these mutants (D635) the aspartic acid (Tf) to serine (p97) switchwhich is evident when comparing the C-terminal regions of Tf and p97 wasput in the N-terminal half of the Tf, and the resulting molecule wasinvestigated. It was found that the max of this mutant Tf was 420 nm (65).For this reason, it is not completely convincing that there is only one atom ofiron bound to the p97 molecule (R. T. A. MacGillivray, personal55communication). It remains to be seen if p97 has one or two functional ironbinding pockets, and if it makes a difference to the function of the molecule.There is an experiment which can be completed using cell linesproduced in this work which might directly address the question of thefunction of p97. It would involve using the parental CHO line TRVB (66),which expresses no functional TR, and the transfectant CHO linesp97aTRVBc3 and p97aTRVBc6, which express high levels of p97. Theexperiment would involve incubating samples of both the transfected anduntransfected TRVB lines with [59Fe] as FeCl3 and then washing the cells andmeasuring the amount of iron internalized. If p97 does indeed contribute tothe iron uptake capacity of the p97aTRVB lines it would follow that theamount of labelled iron internalized would be far greater than for theparental TRVB line. A useful positive control would be the CHO line WTBwhich expresses functional hamster TR, and should take up the iron by thestandard Tf/TR pathway. A further control could be the use of PT-PLC on thep97aTRVB line immediately before the incubation with the labelled iron,which should decrease the amount of iron internalized, if p97 does contributeto iron uptake.One further consequence of the GPI-anchor of p97 is in the potentialcellular distribution of the protein. In polarized epithelial cells, it is knownthat GPI-anchored proteins can be localized to the apical membrane (6 7,68).Presumably, if expressed, p97 will be apically distributed in these cells. TheCaCo-2 cell line has been used as a model system for studying the transport ofiron across epithelial cell monolayers (69). Given that p97 is present in a PTPLC sensitive form on the surface of the CaCo-2 cells, it is possible that thissystem could be used to investigate the potential iron uptake capacity of p97.The experiment would involve treating cells with, or without, bacterial PT-56PLC and observing the effect on iron uptake. If the expression of p97 was toosmall to make a serious contribution to the measurable iron uptake, it wouldbe possible to transfect the CaCo-2 cells with the p97 cDNA in the samemanner as the CHO lines were in this work, to make a system which mightbetter demonstrate the effect. This experiment could extend the work withthe iron uptake in the p97aTRVB cell lines to a more physiological situation.The final functional implication of the GPI-anchor of p97 is that it mayparticipate in signal transduction. There is evidence that GPI-anchoredproteins can be complexed to protein tyrosine kinases, however, it is notknown how these GPI-anchored molecules can transduce signals (70). Thismay be important for the role of p97 in melanomas, and may contribute tocancer progression. It is known that PT turnover is important for vision inDrosophila and cell proliferation in some cultured cells (71). Perhaps themost important part of the p97 molecule is the 1,2 diacylglycerol productswhich result from the reaction catalyzed by PT-PLC, and not the soluble p97released.57CONCLUDING STATEMENT:The major finding in this work is that the human melanomaassociated antigen p97 is attached to the plasma membrane via a GPI-anchor.At the present time investigations are being carried out in our laboratory todetermine both the distribution and the function of the p97 molecule.58REFERENCES:1. Ross, P. M. and D. M. Carter. 1989. Actinic DNA Damage and thePathogenesis of Cutaneous Malignant MelanomaJ. Invest. Dermatol. 92:293s.2. Council on Scientific Affairs., 1989. Harmful Effects of UltravioletRadiation. JAMA 262:380.3. Ananthaswamy, H. N. and W. E. Pierceall. 1990. Molecular Mechanisms ofUltraviolet Radiation Carcinogenesis. Photochem. Photobiol. 52:1119.4. Truhan, A. P. 1991. Sun protection in childhood. Clin. Pediatr. 30:676.5. Robinson, W. A. 1990. Malignant melanoma as a model for cancereducation and prevention. I. Cancer Educ. 5:85.6. Kuhn, L. C., H. M. Schulman, and P. Ponka. 1990. Iron-TransferrinRequirements and Transferrin Receptor Expression in Proliferating Cells. InIron Transport and Storage. P. Ponka, H.M. Schulman and R.C. Woodworth,eds. CRC Press, Boca Raton, p. 149.7. Crichton, R. R. and M. Charloteaux-Wauters. 1987. Iron transport andstorage. Eur. I. Biochem. 164:485.8. Crosa, J. H. 1989. Genetics and Molecular Biology of Siderophore-MediatedIron Transport in Bacteria. Microbiol. Rev. 53 :517.599. Baker, E. N., S. V. Rumball, and B. F. Anderson. 1987. Transferrins: insightsinto structure and function from studies on lactoferrin. Trends Biochem. Sci.12:350.10. Kuhn, L. C. 1989. The transferrin receptor: a key function in ironmetabolism. Schweiz. Med. Wschr. 119:1319.11. Morgan, E. H. 1981. Transferrin, Biochemistry, Physiology and ClinicalSignificance. Molec. Aspects Med. 4:1.12. Basset, P., Y. Quesneau, and J. Zwiller. 1986. Iron-induced L1210 CellGrowth: Evidence of a Transferrin-independent Iron Transport. Cancer Res.46:1644.13. Sturrock, A., J. Alexander, J. Lamb, C. M. Craven, and J. Kaplan. 1990.Characterization of a Transferrin-independent Uptake System for Iron inHeLa Cells. J. Biol. Chem. 265:3139.14. Kaplan, J., I. Jordan, and A. Sturrock. 1991. Regulation of the Transferrinindependent Iron Transport System in Cultured Cells. I. Biol. Chem. 266:2997.15. Thorstensen, K. 1988. Hepatocytes and Reticulocytes Have DifferentMechanisms for the Uptake of Iron from Transferrin. I. Biol. Chern. 263:16837.16. Richardson, D. R. and E. Baker. 1990. The uptake of iron and transferrin bythe human malignant melanoma cell. Biochim. Biophys. Acta 1053:1.6017. Richardson, D. R. and E. Baker. 1991. The release of iron and transferrinfrom the human melanoma cell. Biochim. Biophys. Acta 1091:294.18. Richardson, D. and E. Baker. 1991. The uptake of inorganic iron complexesby human melanoma cells. Biochim. Biophys. Acta 1093:20.19. Richardson, D. R. and E. Baker. 1992. The effect of desferrioxamine andferric ammonium citrate on the uptake of iron by the membrane iron-binding component of human melanoma cells. Biochim. Biophys. Acta1103:275.20. Woodbury, R. G., J. P. Brown, M. Y. Yeh, I. Hellström, and K. E. Hellström.1980. Identification of a cell surface protein, p97’, in human melanomas andcertain other neoplasms. Proc. Nat!. Acad. Sci. USA 77:2183.21. Brown, J. P., R. G. Woodbury, C. E. Hart, I. Hellström, and K. E. Hellström.1981. Quantitative analysis of melanoma-associated antigen p97 in normaland neoplastic tissues. Proc. Nat!. Acad. Sci. USA 78:539.22. Woodbury, R. G., J. P. Brown, S. M. Loop, K. E. Hellström, and I.Hellström. 1981. Analysis of Normal Neoplastic Human Tissues for theTumor-associated Protein p97. mt. j. Cancer 27:145.23. Brown, J. P., K. Nishiyama, I. Hellström, and K. E. Hellström. 1981.Structural Characterization of Human Melanoma-Associated Antigen p97with Monoclonal Antibodies. J. Immunol. 12 7:539.6124. Kahn, M., H. Sugawara, P. McGowan, K. Okuno, S. Nagoya, K. E.Hellström, I. Hellström, and P. Greenberg. 1991. CD4+ T Cell Clones Specificfor the Human p97 Melanoma-Associated Antigen can Eradicate PulmonaryMetastases from a Murine Tumor Expressing the p97 Antigen. I. Immunol.146:3235.25. Brown, J. P., R. M. Hewick, I. Hellström, K. E. Hellström, R. F. Doolittle,and W. J. Dreyer. 1982. Human melanoma-associated antigen p97 isstructurally and functionally related to transferrin. Nature 296:171.26. Plowman, G. D., J. P. Brown, C. A. Erms, J. Schröder, B. Nikinmaa, H. H.Sussman, K. E. Hellström, and I. Hellström. 1983. Assignment of the gene forhuman melanoma-associated antigen p97 to chromosome 3. Nature 303:70.27. Rose, T. M., G. D. Plowman, D. B. Teplow, W. J. Dreyer, K. E. Hellström,and J. P. Brown. 1986. Primary structure of the human melanoma-associatedantigen p97 (melanotransferrin) deduced from the mRNA sequence. Proc.Nat!. Acad. Sci. USA 83:1261.28. Dippold, W. G., K. 0. Lloyd, L. T. C. Li, H. Ikeda, H. F. Oettgen, and L. J. Old.1980. Cell surface antigens of human melanoma: Definition of six antigenicsystems with monoclonal antibodies. Proc. Nati. Acad. Sci. USA 77:6114.29. Liao, S. K., P. C. Kwong, and M. J. Khosravi. 1985. Immunopurification,Characterization, and Nature of Membrane Association of Human62Melanoma-Associated Oncofetal Antigen gp87 Defined by MonoclonalAntibody 140.240. J. Cell. Biochem. 27:303.30. Real, F. X., K. S. Furukawa, M. J. Mattes, S. A. Gusik, C. Cordon-Cardo, H.F. Oettgen, L. J. Old, and K. 0. Lloyd. 1988. Class 1 (unique) tumor antigens ofhuman melanoma: Identification of unique and common epitopes on a 90-kDa glycoprotein. Proc. Nati. Acad. Sci. LISA 85:3965.31. Furukawa, K. S., K. Furukawa, F. X. Real, L. J. Old, and K. 0. Lloyd. 1989. AUnique Antigenic Epitope of Human Melanoma is Carried on the CommonMelanoma Glycoprotein gp95/p97. I. Exp. Med. 169:585.32. Real, F. X., A. N. Houghton, A. P. Albino, C. Cordon-Cardo, M. R.Melamed, H. F. Oettgen, and L. J. Old. 1985. Surface Antigens of Melanomasand Melanocytes Defined by Mouse Monoclonal Antibodies: SpecificityAnalysis and Comparison of Antigen Expression in Cultured Cells andTissues. Cancer Res. 45:4401.33. Sciot, R., R. Dc Vos, P. van Eyken, K. van der Steen, P. Moerman, and V. J.Desmet. 1989. In situ localization of melanotransferrin (melanoma-associatedantigen P97) in human liver. A light- and electronmicroscopicimmunohistochemical study. Liver 9:110.34. Casellas, P., J. P. Brown, 0. Gros, P. Gros, I. Hellström, F. K. Jansen, P.Poncelet, R. Roncucci, H. Vidal, and K. E. Hellström. 1982. Human MelanomaCells can be Killed in vitro by an Immunotoxin Specific for Melanoma-Associated Antigen p97. mt. j. Cancer 30:437.6335. Hu, S. L., G. D. Plowman, P. Sridhar, U. S. Stevenson, J. P. Brown, and C.D. Estin. 1988. Characterization of a Recombinant Vaccinia Virus ExpressingHuman Melanoma-Associated Antigen p97. 1. Virol. 62:176.36. Estin, C. D., U. S. Stevenson, G. D. Plowman, S. L. Hu, P. Sridhar, I.Hellström, J. P. Brown, and K. E. Hellström. 1988. Recombinant vaccinia virusvaccine against the human melanoma antigen p97’ for use inimmunotherapy. Proc. Nat!. Acad. Sci. USA 85:1052.37. Eisenberg, D., E. Schwarz, M. Komaromy, and R. Wall. 1984. Analysis ofMembrane and Surface Protein Sequences with the Hydrophobic MomentPlot. J. Mo!. Biol. 179:125.38. Low, M. G. 1989. The glycosyl-phosphatidylinositol anchor of membraneproteins. Biochim. Biophys. Acta 988:427.39. Lisanti, M. P., E. Rodriguez-Boulan, and A. R. Saltiel. 1990. EmergingFunctional Roles for the Glycosyl-Phosphatidylinositol Membrane ProteinAnchor. J. Membrane Biol. 117:1.40. Ferguson, M. A. J., S. W. Homans, R. A. Dwek, and T. W. Rademacher.1988. Glycosyl-Phosphatidylinositol Moiety That Anchors Trypanosomabrucei Variant Surface Glycoprotein to the Membrane. Science 23 9:753.41. Homans, S. W., M. A. J. Ferguson, R. A. Dwek, T. W. Rademacher, R.Anand, and A. F. Williams. 1988. Complete structure of the glycosyl64phosphatidylinositol membrane anchor of rat brain Thy-i glycoprotein.Nature 333:269.42. Doering, T. L., W. J. Masterson, G. W. Hart, and P. T. Englund. i990.Biosynthesis of Glycosyl Phosphatidylinositol Membrane Anchors. I. Biol.Chem. 265:6ii.43. Moran, P. and I. W. Caras. 1991. A Nonfunctional Sequence Converted to aSignal for Glycophosphatidylinositol Membrane Anchor Attachment. 1. CellBiol. 115:329.44. Moran, P. and I. W. Caras. 1991. Fusion of Sequence Elements from Non-anchored Proteins to Generate a Fully Functional Signal forGlycophosphatidylinositol Membrane Anchor Attachment. 1. Cell Biol.115:1595.45. Keller, G. A., M. W. Siegel, and I. W. Caras. 1992. Endocytosis ofglycophospholipid-anchored and transmembrane forms of CD4 by differentendocytic pathways. EMBO 111:863.46. Trowbridge, I. S., R. Hyman, and C. Mazauskas. 1978. The Synthesis andProperties of T25 Glycoprotein in Thy-i-Negative Mutant Lymphoma Cells.Cell 14:21.47. Ferguson, M. A. J. and A. F. Williams. 1988. Cell-Surface Anchoring OfProteins Via Glycosyl-Phosphatidylinositol Structures. Ann. Rev. Biochem.57:285.6548. Metz, C. N., Y. Zhang, Y. Guo, T. C. Tsang, J. P. Kochan, N. Altszuler, andM. A. Davitz. 1991. Production of the Glycosyiphosphatidylinositol-specificPhospholipase D by the Islets of Langerhans. I. Biol. Chem. 266:17733.49. Peiham, H. R. B. 1991. Multiple Targets for Brefeldin A. Cell 67:449.50. Klausner, R. D., J. G. Donaldson, and J. Lippincott-Schwartz. 1992.Brefeldin A: Insights into the Control of Membrane Traffic and OrganelleStructure. J. Cell. Bid. 116:1071.51. Lippé, R., E. Luke, Y. T. Kuah, C. Lomas, and W. A. Jefferies. 1991.Adenovirus Infection Inhibits the Phosphorylation of MajorHistocompatibility Complex Class I Proteins. I. Exp. Med. 174:1159.52. Low, M. G., J. Stiernberg, G. L. Waneck, R. A. Flavell, and P. W. Kincade.1988. Cell-specific heterogeneity in sensitivity of phosphatidylinositolanchored membrane antigens to release by phospholipase C. I. Immunol.Methods 113:101.53. von Boxberg, Y., R. Wütz, and U. Schwarz. 1990. Use of the biotin-avidinsystem for labelling, isolation and characterization of neural cell-surfaceproteins. Eur. I. Biochem. 190:249.54. Bordier, C. 1981. Phase Separation of Integral Membrane Proteins in TritonX-114 Solution. J. Biol. Chem. 256:1604.6655. Estin, C. D., U. Stevenson, M. Kahn, I. Hellström, and K. E. Hellström.1989. Transfected Mouse Melanoma Lines That Express Various Levels ofHuman Melanoma-Associated Antigen p97. 1. Natl. Cancer Inst. 81:445.56. Low, M. G. and P. W. Kincade. 1985. Phosphatidylinositol is themembrane-anchoring domain of the Thy-i glycoprotein. Nature 3 18:62.57. Takami, N., K. Oda, T. Fujiwara, and Y. Ikehara. 1990. Intracellularaccumulation and oligosaccharide processing of alkaline phosphatase underdisassembly of the Golgi complex caused by brefeldin A. Eur. I. Biochem.194:805.58. Urbani, L. and R. D. Simoni. 1990. Cholesterol and Vesicular StomatitisVirus G Protein Take Separate Routes from the Endoplasmic Reticulum tothe Plasma Membrane. J. Biol. Chem. 265:1919.59. Vance, J. E., E. J. Aasman, and R. Szarka. 1991. Brefeldin A Does NotInhibit the Movement of Phosphatidylethanolamine from Its Sites ofSynthesis to the Cell Surface. I. Biol. Chem. 266:8241.60. Rothberg, K. G., Y. Ying, J. F. Koihouse, B. A. Kamen, and R. G. W.Anderson. 1990. The Glycophospholipid-linked Folate Receptor InternalizesFolate Without Entering the Clathrin-coated Pit Endocytic Pathway. I. CellBiol. 110:637.6761. Anderson, R. G. W., B. A. Kamen, K. G. Rothberg, and S. W. Lacey. 1992.Potocytosis: Sequestration and Transport of Small Molecules by Caveolae.Science 255:410.62. Rothberg, K. G., J. E. Heuser, W. C. Donzell, Y. S. Ying, J. R. GlemLey, andR. G. W. Anderson. 1992. Caveolin, a Protein Component of CaveolaeMembrane Coats. Cell 68:673.63. Baker, E. N., H. M. Baker, C. A. Smith, M. R. Stebbins, M. Kahn, K. E.Hellström, and I. Hellström. 1992. Human melanotransferrin (p97) has onlyone functional iron-binding site. FEBS LETTERS 298:215.64. Anderson, B. F., H. M. Baker, G. E. Norris, S. V. Rumball, and E. N. Baker.1990. Apolactoferrin structure demonstrates ligand-induced conformationalchange in transferrins. Nature 344:784.65. Woodworth, R. C., A. B. Mason, W. D. Funk, and R. T. A. MacGillivray.1991. Expression and Initial Characterization of Five Site-Directed Mutants ofthe N-Terminal Half-Molecule of Human Transferrin. Biochemistry 30:10824.66. McGraw, T. E., L. Greenfield, and F. R. Maxfield. 1987. FunctionalExpression of the Human Transferrin Receptor cDNA in Chinese HamsterOvary Cells Deficient in Endogenous Transferrin Receptor. I. Cell Biol.105:207.67. Lisanti, M. P., A. Le Bivic, A. R. Saltiel, and E. Rodriguez-Boulan. 1990.Preferred Apical Distribution of Glycosyl-Phosphatidylinositol (GPI)68Anchored Proteins: A Highly Conserved Feature of the Polarized EpithelialCell Phenotype. I. Membrane Biol. 113:155.68. Lisanti, M. P., I. W. Caras, M. A. Davitz, and E. Rodriguez-Boulan. 1989. AGlycophospholipid Membrane Anchor Acts as an Apical Targetting Signal inPolarized Epithelial Cells. I. Cell Biol. 109:2145.69. Alvarez-Hernandez, X., G. M. Nichols, and J. Glass. 1991. Caco-2 cell line: asystem for studying intestinal iron transport across epithelial cell monolayers.Biochim. Biophys. Acta 1070:205.70. Stefanová, I., V. HoejI, I. J. Ansotegui, W. Knapp, and H. Stockinger.1991. GPI-Anchored Cell-Surface Molecules Complexed to Protein TyrosineKinases. Science 254:1016.71. Majerus, P. W., T. S. Ross, T. W. Cunningham, K. K. Caidwell, A. B.Jefferson, and V. S. Bansal. 1990. Recent Insights in PhosphatidylinositolSignaling. Cell 63:459.


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items